“A Warning is Not Enough. Teach Me How to Spot Deepfakes.”: Testing Media Literacy Interventions for Combating Deepfakes

The rise of generative artificial intelligence (AI) has supercharged the proliferation of multimodal disinformation or so-called “deepfakes.” Through a two-wave online experiment among U.S. residents, we tested media literacy interventions that varied by message content (warning, technique tips, combined) and media format (text, infographic, video) on their effectiveness in combating deepfakes. Findings showed technique tips on how to identify deepfakes improved deepfake detection, perceived AI literacy, and self-efficacy, while the video format was most effective in elevating perceived AI literacy. One week later, the detection accuracies remained consistently high across the board, but perceived AI literacy and self-efficacy decreased in some conditions.